PROCEEDINGS OF THE 31st ICRC, ŁÓDŹ 2009 1

Search for Exotic Physics with the ANTARES Detector

Gabriela Pavalas∗, Nicolas Picot Clemente†

on behalf of the ANTARES Collaboration

∗Institute for Space Sciences, Bucharest-Magurele, Romania†CNRS / Centre de Physique des Particules de Marseille, Marseille, France

Abstract. Besides the detection of high energyneutrinos, the ANTARES telescope offers an oppor-tunity to improve sensitivity to exotic cosmologicalrelics. In this article we discuss the sensitivity ofthe ANTARES detector to relativistic monopoles andslow nuclearites. Dedicated trigger algorithms andsearch strategies are being developed to search forthem. The data filtering, background rejection selec-tion criteria are described, as well as the expectedsensitivity of ANTARES to exotic physics.

Keywords: ANTARES, magnetic monopoles, nucle-arites

I. I NTRODUCTION

The ANTARES neutrino telescope is aimed to observehigh energy cosmic neutrinos through the detectionof the Cherenkov light produced by up-going inducedmuons. However, the ANTARES detector is also sen-sitive to a variety of exotic particles, and can providean unique facility for the search of magnetic monopolesand nuclearites.

II. T HE ANTARES DETECTOR

The ANTARES detector has reached its nominalsize in May 2008. The 884 Optical Modules (OM)are deployed on twelve vertical lines in the WesternMediterranean, at depths between 2050 and 2400 meters.The OMs, consisting of a glass sphere housing a 10”Hamamatsu photomultiplier (PMT) [1], are arrangedby triplet per storey. Each detector line, made of 25storeys, is connected via interlinks to a Junction Box,itself connected to the shore station at La Seyne-sur-Mer through a 40 km long electro-optical cable. Thestrategy of the ANTARES data acquisition is based onthe “all-data-to-shore” concept [2]. This implementationleads to the transmission of all raw data above a giventhreshold to shore, where different triggers are appliedto the data for their filtering before storage.For the analysis presented here, only the two generaltrigger logics operated up to now have been considered.Both are based on local coincidences. A local coinci-dence (L1 hit) is defined either as a combination oftwo hits on two OMs of the same storey within 20ns, or as a single hit with a large amplitude, typically3 pe. The first trigger, a so-called directional trigger,requires five local coincidences anywhere in the detectorbut causally connected, within a time window of 2.2µs.The second trigger, a so-called cluster trigger, requires

two T3-clusters within 2.2µs, where a T3-cluster isa combination of two L1 hits in adjacent or next-to-adjacent storeys. When an event is triggered, all PMTpulses are recorded over 2.2µs.The ANTARES observatory was build gradually, givingrise to various detector layouts used for physics analysis.The 5-line, 10-line and 12-line detector configurationsmatch with data taken from January 2007, from Decem-ber 2007 and from May 2008, respectively.

III. M AGNETIC MONOPOLES

A. Introduction

Most of the Grand Unified Theories (GUTs) pre-dict the creation of magnetic monopoles in the earlyUniverse. Indeed, in 1974, ’t Hooft [3] and Polyakov[4] showed independently that each time a compactand connected gauge group is broken into a connectedsubgroup, elements caracterising well a magnetic charge,as introduced by Dirac in 1931 [5], appear.These particles are topologically stable and carry amagnetic charge defined as a multiple integer of theDirac chargegD = h̄c2e , where e is the elementary electriccharge, c the speed of light in vacuum andh̄ the Planckconstant. Depending on the group, the masses inferredfor magnetic monopoles can take range over many ordersof magnitude, from108 to 1017 GeV.

As magnetic monopoles are stable, and so would sur-vive until now, they should have been very diluted in theUniverse, as predicted by numerous theoretical studieswhich set stringent limit on their fluxes, like the Parkerflux limit [6]. More stringent limits were set recently bydifferent experiments (MACRO [7], AMANDA [8], ...).

The development of neutrino astronomy in the lastdecade led to the construction of huge detectors, whichallow new hopes in the search for magnetic monopoles.Actually, the ANTARES detector seems to be welldesigned to detect magnetic monopoles, or at least toimprove limits on their fluxes, as described below.

B. Signal and background simulations

Since fast monopoles have a large interaction withmatter, they can lose large amounts of energy in theterrestrial environment. The total energy loss of a rela-tivistic monopole with one Dirac charge is of the orderof 1011 GeV [9] after having crossed the full diameterof the Earth. Because magnetic monopoles are expectedto be accelerated in the galactic coherent magnetic fielddomain to energies of about1015 GeV [10], some could

2 G. PAVALAS et al. EXOTIC PHYSICS WITH ANTARES

be able to cross the Earth and reach the ANTARESdetector as upgoing signals.

The monopole’s magnetic chargeg = ngD can beexpressed as an equivalent electric chargeg = 68.5ne,where n is an integer. Thus relativistic monopoles withβ ≥ 0.74 carrying one Dirac charge will emit a largeamount of direct Cherenkov light when traveling throughthe ANTARES detector, giving rise to∼ 8500 moreintense light than a muon. The number of photons perunit length (cm−1) emitted on the path of a monopoleis shown in figure 1, as a function of the velocity of themonopole up toγ = 10 (β = 0.995).

Fig. 1. Number of emitted photons per unit length (cm−1) by amagnetic monopole with a chargeg = gD through direct Cherenkovemission (dashed line) compared to the number of photons emittedbya muon (black line), as a function of their velocities.

For the analysis, monopoles have been simulatedinside an optimized volume containing the 12-line de-tector, for six ranges of velocities betweenβ = 0.74 andβ = 0.995. In addition, downgoing atmospheric muonshave been simulated using the CORSIKA package [11],as well as upgoing and downgoing atmospheric neu-trinos according to the Bartol flux [12], [13]. Opticalbackground from40K decay has been added to bothmagnetic monopole signal and atmospheric muon andneutrino background events.

C. Search strategy

The 12-line detector data are triggered by both trig-ger logics, the directional and the cluster triggers (seesection II). A comparison of efficiency was thereforeperformed on magnetic monopoles and restricted only toupward-going events. As the efficiency of the directionaltrigger was found to be lower than for the cluster trigger,it was decided to perform searches for upgoing magneticmonopoles with the cluster trigger only.

The standard reconstruction algorithm, developped inANTARES for upward-going neutrino selection, andmainly based on a likelihood maximization, was applied.

In order to select upgoing particles, only reconstructedevents with a zenith angle lower than 90◦ were selected.However, muon bundles are difficult to reconstruct pro-perly and some of them can be reconstructed as upward-going events.As it is shown in figure 2, a magnetic monopole tra-versing the detector will emit an impressive quantity oflight, compared to atmospheric muons or muons inducedby atmospheric neutrinos. The large amount of induced

Fig. 2. Normalized events as a function of the number of T3 clustersfor downgoing atmospheric muons, upgoing and downgoing atmo-spheric neutrinos, and upgoing magnetic monopoles withβ ∼ 0.75andβ ∼ 0.99.

hits in the detector, more precisely the number of T3clusters, is therefore used as a criteria to remove part ofthe atmospheric background.

Before applying a supplementary cut to reduce theremaining background, 10 active days of golden1 datawere taken as reference to check the data MonteCarloagreement. The comparison of T3 distributions betweenthe data and the background simulation for 10 days isshown in figure 3.

Fig. 3. Comparison of T3 distributions between data and MonteCarlosimulations for 10 days of data taking.

We optimized the cuts on the number of T3 clustersto maximize the 90% C.L. sensitivity, calculated withthe usual Feldman-Cousins formula[14], for magneticmonopoles after one year of data taking. In the opti-mization process the same selection criteria have been

1Golden data assumes experimental data complying with certainselection criteria like low baserate and burstfraction.

PROCEEDINGS OF THE 31st ICRC, ŁÓDŹ 2009 3

applied to calculate the sensitivity to magnetic monopoleevents over the whole velocity range0.74 ≤ β ≤ 0.995.

Finally the 90% C.L. sensitivity for this range wasfound, for a cut of at least 140 T3 clusters, for whichabout 1.7 background events are expected. The 90% C.L.sensitivity for ANTARES after one year of data takingis of the order of∼ 1 · 10−17cm−2s−1sr−1.

IV. N UCLEARITES

A. Introduction

Nuclearites are hypothetical nuggets of strange quarkmatter that could be present in cosmic radiation.Their origin is related to energetic astrophysical pheno-mena (supernovae, collapsing binary strange stars, etc.).Down-going nuclearites could reach the ANTARESdepth with velocities∼ 300 km/s, emitting blackbodyradiation at visible wavelengths while traversing seawater.

Heavy nuggets of strange quark matter (M ≥1010GeV), known as nuclearites, would be electricallyneutral; the small positive electric charge of the quarkcore would be neutralized by electrons forming anelectronic cloud or found in week equilibrium inside thecore. The relevant energy loss mechanism is representedby the elastic collisions with the atoms of the traversedmedia, as shown in ref. [16]:

dE

dx= −σρv2, (1)

whereρ is the density of the medium,v is the nuclearitevelocity andσ its geometrical cross section:

σ =

{

π(3M/4πρN )2/3 for M ≥ 8.4 ∗ 1014 GeV;

π × 10−16cm2 for lower masses.

with ρN = 3.6 × 1014 g cm−3. The mass limit in theabove equation corresponds to a radius of the strangequark matter of 1̊A. Assuming a nuclearite of massMenters the atmosphere with an initial (non-relativistic)velocity v0, after crossing a depth L it will be sloweddown to

v(L) = v0e−

σ

M

∫

L

0

ρdx (2)

whereρ is the air density at different depths. Consideringthe parametrization of the standard atmosphere from[17]:

ρ(h) = ae−h

b = ae−H−l

b , (3)

where a=1.2 × 10−3 g cm−3 and b≃ 8.57 × 105 cm,H≃50 km is the total height of the atmosphere, theintegral in Eq.2 may be solved analytically.

The propagation of nuclearites in sea water is de-scribed also by Eq. 2, assumingρ = 1 g cm−3 andsubstitutingv0 with the speed value at the Earth surface.Nuclearites moving into the water could be detectedbecause of the black-body radiation emitted by theexpanding cylindrical thermal shock wave [16]. Theluminous efficiency (defined as the fraction of dissipated

Entries 2293

number of snapshots0 20 40 60 80 100 120

eve

nts

1

10

210

Entries 2293

Fig. 4. Snapshot distribution obtained for simulated nuclearite eventswith masses 3×1016, 1017 and 1018 GeV.

energy appearing as light) was estimated, in the case ofwater, to beη ≃ 3 × 10−5 [16]. The number of visiblephotons emitted per unit path length can be computedas follows:

dNγdx

= ηdE/dx

π(eV ), (4)

assuming the average energy of visible photonsπ eV.

B. Search strategy and results

The Monte Carlo simulation of nuclearite detectionin ANTARES assumes only the down-going part ofan isotropic flux of nuclearites and an initial velocity(before entering the atmosphere) ofβ = 10−3. A typicalnuclearite event would cross the Antares detector in acharacteristic time of∼ 1 ms, producing a luminositythat would exceed that of muons by several orders ofmagnitude. Down-going atmospheric muons representthe main background for nuclearite events. This analysisrefers to simulated nuclearite and muon events using the5-line detector configuration and 84 days of data takenfrom June to November 2007.Simulated nuclearite and muon events have been pro-cessed with the directional trigger, that operated duringthe 5-line data acquisition. Background was added froma run taken in July 2007, at a baserate of 63.5 kHz. Wefound for nuclearites a lower mass limit detectable withthe directional trigger of3×1016 GeV. Nuclearite eventswere simulated for masses of3 × 1016 GeV, 1017 GeVand1018 GeV. The atmospheric muons were generatedwith the MUPAGE code [18]. The parameters used inour analysis comprised the number of L1 triggered hits(see Section II), the number of single hits (L0 hits,defined as hits with a threshold greater than 0.3 pho-toelectrons), the duration of the snapshot (defined as thetime difference between the last and the first L1 triggeredhits of the event) and total amplitude of hits in theevent. Data were reprocesed with the directional trigger.We obtained a good agreement between simulated muonevents and data.

The algorithm of the directional trigger selects fromall the hits produced by a nuclearite only those that

4 G. PAVALAS et al. EXOTIC PHYSICS WITH ANTARES

Entries 47606Mean 484RMS 253.4

dt (ns)0 2000 4000 6000 8000 10000 12000 14000

norm

aliz

ed

-410

-310

-210

-110 Entries 47606Mean 484RMS 253.4

nuclearites

muons

data

Fig. 5. Normalized distributions as a function of the duration ofsnapshot. Comparison between data (continuous line), simulated muon(dotted line) and nuclearite (dashed line) events is shown.

comply to the signal of a relativistic muon. Thesehits can be contained in a single snapshot or multiplesnaphots for a single event.

Because nuclearites are slowly moving particles, mul-tiple snapshots belonging to a single nuclearite eventmay span into intervals from tens ofµs up to∼1ms.The multiplicity of snapshots in the simulated nucle-arite events is presented in Fig. 4. The majority ofthe snapshots produced by nuclearites from the studiedsample are of short duration (